Episode 282 - Noblis-FBI Black Box 2 Fingerprint Study

Double Loop Podcast

In this episode, Glenn and Eric start with a quick game of “A Truth, A Lie, and a Mandella Effect”. Then they pose a question from a listener regarding interest in a forum where listeners can go and discuss the episodes of the show, ask questions, and chat “amongst themselves”. Finally they get to the long awaited Noblis/FBI study “Black Box 2” latent fingerprint error rate study. This is a re-do of the 2011 black box study for latent fingerprint performance. This study “Accuracy and reproducibility of latent print decisions on comparisons from searches of an automated fingerprint identification system” by Hicklin, Richetelli, Taylor and Buscaglia (For Sci Intl, 370 (2025), 112457), reported the performance of 156 U.S. latent print examiner participants each reviewing about 100 latent prints/comparisons resulting in over 14,000 trials. The study reports numerous statistics of performance such as sensitivity (63%), specificity (70%), false positive error rate (0.2%), and false negative error rate (4.2%). The guys also discuss some important study design differences, plus there are a lot of data in the appendices. Finally they talk about the number of false positives made in the study (n=23) and the fact that 13 of those 23 were made by one participant and how that impacted the results. Article is available for free at: https://doi.org/10.1016/j.forsciint.2025.112457